領先指標對景氣循環之驗證

An Estimation of Leading Indicators and Business Cycle

綦茵蘋(Yin-Ping Chi)

指導教授 ： 林建甫

國立臺灣大學/社會科學院/經濟學研究所/碩士(2007年)

https://doi.org/10.6342/NTU.2007.01301

摘要

領先指數含有領先景氣循環的特點，常被廣泛的用來預測或監控經濟波動，或用來預測未來的整體經濟活動，屬於重要的景氣預測指標，行政院經濟建設委員會參酌美國國家經濟研究局(National Bureau of Economic Research, NBER)的編製方法，自1977年開始按月公佈台灣景氣動向指標，由於該指標已延用約30年且無重大更動，對於領先指標是否仍適用於現今的經濟情勢，或是否仍具有領先的意義，是本研究主要的目的，另外並試圖尋找並編製更能符合領先意義的指標。
由於目前經建會所編製之領先指標，雖然經圖形驗證及峰谷對應法檢視仍具領先意義，但經是否能以計量方法檢驗仍然值得探討，本研究利用單根檢定、共整合、因果關係檢定及誤差修正模型等時間數列之統計分析方法，檢驗經濟成長與同時指標、領先指標之關係，並分析現行領先指標之構成項目與經濟活動之關連性，最後則選取更能代表景氣循環的變數以構成新的領先指標。經實證結果發現，現行同時指標與經濟成長變數具有高度相關性，且同時指標與實質GDP具有共整關係，表示二變數間具有長期均衡的關係。
而現行領先指標大致能補捉到經濟循環的波動，但領先的效果卻不甚明顯，在7項構成項目與實質GDP的分析結果，雖發現具有長期均衡的關係，但台灣區房屋建築申請面積、製造業新接訂單指數、躉售物價指數及股價指數等4個項目的領先關係均未見明顯。故保留原有構成項目中的貨幣供給M1B、海關出口值及製造業平均工時，另加上外銷訂單總額、CPI指數、美國領先指標及NASDAQ指數4項經濟數據，替換原有不顯著之構成項目而編製成新的領先指標，在相關係數、因果關係檢定及線性廻歸方面，均較現行領先指標對經濟成長率有顯著的領先效果。

關鍵字

領先指標 ； 單根檢定 ； 因果關係檢定 ； 誤差修正模型

並列摘要

Leading indicators, which signal future business cycles, are widely used to predict or monitor the fluctuations of economic activities. They are also designed to provide early signs of economic trends and, therefore, considered significant economic predictive tools.
Using the compilation of the US’s National Bureau of Economic Research as a reference, the Council for Economic Planning and Development (CEPD) in Taiwan has been announcing the nation’s business indicators on a monthly basis since 1977. However, the CEPD made no revisions to the compilation of its leading indicator in the past 30 years, which thus raises doubts of whether the indicator is outdated and how effectively it can forecast economic trends. This study not only aims to answer the doubts, but also attempts to find and compile a new leading indicator that better reflects future economic turns.
The CEPD’s leading indicator still shows significance after going through graphic verification and peak-valley detection. But the question of whether it is as significant under the econometrics remains in doubt. Therefore, this study will utilize statistic analyzing methods such as the unit root test, cointegration, Granger causality test and error correction model (ECM) to examine the relationship among the economic growth, coincident indicators and leading indicators. The correlation between components of a leading indicator and economic activities will be analyzed as well to come up with a new leading indicator that better fore-tells business cycles.
The empirical evidence showed that coincident indicators and variables of economic growth are highly correlated. There also exists a cointegration between coincident indicators and the GDP, which shows that a long-term equilibrium relationship has long been present.
The current leading indicator is capable of capturing most of the economic turns, but is less obvious in predicting. There also exists a long-term equilibrium relationship between seven of the indicator’s components and the GDP. However, among the seven components, four components including the floor area for new housing projects in Taiwan, new orders for manufactured goods, the wholesale index and stock price index are not significant. Therefore, three of the indicator’s original components including the monetary aggregate M1B, exports by customs and the manufacturing sector’s average monthly working hours are retained while four new components including export orders, the consumer price index (CPI), the US’s leading indicator and the NASDAQ composite index are replaced to make up a new leading indicator. The new indicator has shown great significance in leading as far as correlation co-efficiency, Granger causality test and the linear regression model are concerned.

並列關鍵字

leading indicator、unit root test ； Granger causality test ； error correction model

參考文獻

13. 梁國源、高志祥、周大森(2005)，臺灣與美國跨國景氣互動之分析－從產業關聯層探討，《臺灣經濟預測與政策》，35:2，p43-78。

15. 陳仕偉、沈中華(2003)，金融領先指標與實質領先指標訊息一致嗎？台灣領先指標的實證分析，《人文及社會科學集刊》15:4，p627-660。

18. 彭素玲、周濟(2001)，臺灣總體經濟即期季模型之建立與應用，《台灣經濟預測與政策》32:1，p77-116。

9. Lin, J.L. and Tian-Shy, L., (2003), "Modeling Lunar Calendar Holiday Effects in Taiwan", Taiwan Economic Policy and Forecast, 33,2,1-37

3. David, F.H(1980), "Econometrics-alchemy or science? ", Economica, 47, 188, 387-406